Browse Prior Art Database

Method to Refine Image Selection and Update for Enhancing Augmented Reality Service

IP.com Disclosure Number: IPCOM000245619D
Publication Date: 2016-Mar-22
Document File: 3 page(s) / 63K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method of multi-factor based image management to refine the best-fitting image candidate selection for enhancing augmented reality service.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 39% of the total text.

Page 01 of 3

Method to Refine Image Selection and Update for Enhancing Augmented Reality Service

Augmented Reality (AR), Virtual Reality (VR), and related Internet of Things (IoT) services are growing areas in computing research and the information technology (IT) markets. Many IoT-related services rely on real-time camera input features to gather real world images around users as primary contextual and spatial data, and then use that information to provide better IoT service.

One current IoT service uses a user's smart phone camera as an initial input device for an Optical Character Recognition (OCR)-based Machine Translation Service (OCRMTS). OCRMTS is an image-based translation tool and is one of the most important machine translation features for globalization support and services. The service translates information content from the first language on image A to the second language on image-B.

One cloud recognition service takes the user's smart phone camera input for an augmented reality service. In addition, on mobile and wearable devices, voice, text, and image translation based machine translation applications are widely used. OCR machine translation (e.g., camera input image MT) has been widely used to help people read foreign text information on billboards, street signs, or other information boards in different countries/regions.

For an Augmented Reality Service, camera input is the first step to sending service requests. Those services require high-speed networks with better bandwidth. Network latency and security restrictions may be major problems and limitations for receiving better online AR service. Working offline is not an optional for any IoT services. For instance, an Online Camera Image MT Service (CIMTS) requests users to send a source image with foreign language text to a centralized CIMTS server for getting services of OCR conversion, Natural Language Processing (NLP) analysis, machine translation process etc. Therefore, the online CIMTS needs greater Internet bandwidth and a more powerful Central Processing Unit (CPU) on the server than any other online services due to OCR and NLP processes.

An off-line CIMTS is another option to solve the latency problem, but it normally needs 10-20 times more storage space on the client side and a more powerful computing architecture to support intensive computing tasks for supporting services of OCR conversion, NPL analysis, and normal machine translation processes. This is not an option for wearable/mobile and other IoT solutions.

The above technologies rely on either the user's real time inputs or mange images at a single level. There are no ideal methods to evaluate, manage, identify, compare, and catalog those user inputs for dimensions of purpose, subject, object, location, time, and any contextual factors.

Using a similar image collected from other users is a good way to solve the above

1


Page 02 of 3

problems. Refining the collected images and providing the best-fitting image...